Variability in clinical classification of raw test scores across normativedata sets

Sampling differences and other sources of measurement error result in varia tion between normative data sets reported for the same test; however, no re search has examined the degree to which classification of a specific score may be affected by discrepancies across norms. The present study applies di fferent sets of published normative data to test protocols of 12 men and wo men of various ages and levels of education. The tests included in the fict itious assessment battery are used routinely by neuropsychologists accordin g to Butler, Retzlaff, and Vanderploeg (1991) and have at least two sets of published norms. Our findings indicate that interpretation of performance on neuropsychological tests is frequently and dramatically affected by samp ling differences across normative data sets. Suggestions for the selection of normative data sets for clinical assessments and the development of comp rehensive normative data sets are offered.